Team Ai
Datasetpublic

codekingpro/portable-devtools

sourceHugging Faceupdated 5mo agoView on Hugging Face
1likes15kdownloads
_exceptions.py163 linesDownload Raw Back to _core
1"""
2Various richly-typed exceptions, that also help us deal with string formatting
3in python where it's easier.
4
5By putting the formatting in `__str__`, we also avoid paying the cost for
6users who silence the exceptions.
7"""
8
9def _unpack_tuple(tup):
10    if len(tup) == 1:
11        return tup[0]
12    else:
13        return tup
14
15
16def _display_as_base(cls):
17    """
18    A decorator that makes an exception class look like its base.
19
20    We use this to hide subclasses that are implementation details - the user
21    should catch the base type, which is what the traceback will show them.
22
23    Classes decorated with this decorator are subject to removal without a
24    deprecation warning.
25    """
26    assert issubclass(cls, Exception)
27    cls.__name__ = cls.__base__.__name__
28    return cls
29
30
31class UFuncTypeError(TypeError):
32    """ Base class for all ufunc exceptions """
33    def __init__(self, ufunc):
34        self.ufunc = ufunc
35
36
37@_display_as_base
38class _UFuncNoLoopError(UFuncTypeError):
39    """ Thrown when a ufunc loop cannot be found """
40    def __init__(self, ufunc, dtypes):
41        super().__init__(ufunc)
42        self.dtypes = tuple(dtypes)
43
44    def __str__(self):
45        return (
46            f"ufunc {self.ufunc.__name__!r} did not contain a loop with signature "
47            f"matching types {_unpack_tuple(self.dtypes[:self.ufunc.nin])!r} "
48            f"-> {_unpack_tuple(self.dtypes[self.ufunc.nin:])!r}"
49        )
50
51
52@_display_as_base
53class _UFuncBinaryResolutionError(_UFuncNoLoopError):
54    """ Thrown when a binary resolution fails """
55    def __init__(self, ufunc, dtypes):
56        super().__init__(ufunc, dtypes)
57        assert len(self.dtypes) == 2
58
59    def __str__(self):
60        return (
61            "ufunc {!r} cannot use operands with types {!r} and {!r}"
62        ).format(
63            self.ufunc.__name__, *self.dtypes
64        )
65
66
67@_display_as_base
68class _UFuncCastingError(UFuncTypeError):
69    def __init__(self, ufunc, casting, from_, to):
70        super().__init__(ufunc)
71        self.casting = casting
72        self.from_ = from_
73        self.to = to
74
75
76@_display_as_base
77class _UFuncInputCastingError(_UFuncCastingError):
78    """ Thrown when a ufunc input cannot be casted """
79    def __init__(self, ufunc, casting, from_, to, i):
80        super().__init__(ufunc, casting, from_, to)
81        self.in_i = i
82
83    def __str__(self):
84        # only show the number if more than one input exists
85        i_str = f"{self.in_i} " if self.ufunc.nin != 1 else ""
86        return (
87            f"Cannot cast ufunc {self.ufunc.__name__!r} input {i_str}from "
88            f"{self.from_!r} to {self.to!r} with casting rule {self.casting!r}"
89        )
90
91
92@_display_as_base
93class _UFuncOutputCastingError(_UFuncCastingError):
94    """ Thrown when a ufunc output cannot be casted """
95    def __init__(self, ufunc, casting, from_, to, i):
96        super().__init__(ufunc, casting, from_, to)
97        self.out_i = i
98
99    def __str__(self):
100        # only show the number if more than one output exists
101        i_str = f"{self.out_i} " if self.ufunc.nout != 1 else ""
102        return (
103            f"Cannot cast ufunc {self.ufunc.__name__!r} output {i_str}from "
104            f"{self.from_!r} to {self.to!r} with casting rule {self.casting!r}"
105        )
106
107
108@_display_as_base
109class _ArrayMemoryError(MemoryError):
110    """ Thrown when an array cannot be allocated"""
111    def __init__(self, shape, dtype):
112        self.shape = shape
113        self.dtype = dtype
114
115    @property
116    def _total_size(self):
117        num_bytes = self.dtype.itemsize
118        for dim in self.shape:
119            num_bytes *= dim
120        return num_bytes
121
122    @staticmethod
123    def _size_to_string(num_bytes):
124        """ Convert a number of bytes into a binary size string """
125
126        # https://en.wikipedia.org/wiki/Binary_prefix
127        LOG2_STEP = 10
128        STEP = 1024
129        units = ['bytes', 'KiB', 'MiB', 'GiB', 'TiB', 'PiB', 'EiB']
130
131        unit_i = max(num_bytes.bit_length() - 1, 1) // LOG2_STEP
132        unit_val = 1 << (unit_i * LOG2_STEP)
133        n_units = num_bytes / unit_val
134        del unit_val
135
136        # ensure we pick a unit that is correct after rounding
137        if round(n_units) == STEP:
138            unit_i += 1
139            n_units /= STEP
140
141        # deal with sizes so large that we don't have units for them
142        if unit_i >= len(units):
143            new_unit_i = len(units) - 1
144            n_units *= 1 << ((unit_i - new_unit_i) * LOG2_STEP)
145            unit_i = new_unit_i
146
147        unit_name = units[unit_i]
148        # format with a sensible number of digits
149        if unit_i == 0:
150            # no decimal point on bytes
151            return f'{n_units:.0f} {unit_name}'
152        elif round(n_units) < 1000:
153            # 3 significant figures, if none are dropped to the left of the .
154            return f'{n_units:#.3g} {unit_name}'
155        else:
156            # just give all the digits otherwise
157            return f'{n_units:#.0f} {unit_name}'
158
159    def __str__(self):
160        size_str = self._size_to_string(self._total_size)
161        return (f"Unable to allocate {size_str} for an array with shape "
162                f"{self.shape} and data type {self.dtype}")
163 
codekingpro/portable-devtools · Team Ai